At Adaptive, I helped to benchmark the Aeron messaging platform on Amazon Web Services (AWS) and Google Cloud Platform (GCP).

During the course of the project, I:

  • built the infrastructure provisioning harness required to run the tests on two cloud providers with different scenarios and configurations
    • instance types
    • network topology and configuration
    • transport security (IPSec, Wireguard, etc.)
    • storage types
  • built an automation for rendering results (latency graphs, experiment parameters, etc.)
  • built an automation for provisioning infrastructure, running benchmarks in various configurations (peer-to-peer, cluster) and rendering results
  • assisted with the running of benchmarks, analyzing results and investigating performance issues

See the press release from AWS on the topic and the performance testing results are publicly available.

See also the article on achieving sub-20 microsecond performance on Google Cloud.

Tools used: AWS, GCP, Linux, Terraform, Bash, Performance Engineering, Python, Java